RAdio Detection And Ranging, often referred to as Radar, is a technology that has been increasingly used in many vehicular applications, such as in adaptive cruise control, sensor-based applications, etc. A radar system is an electronic system designed to transmit radio signals and receive reflected images of those signals from a ‘target’ object, in order to determine the bearing and distance to the ‘target’. In future, vehicle manufacturers have suggested that vehicle radar systems may additionally be employed in safety related applications, such as: determination of a vehicle starting and/or stopping, to perform pre-cash detection and potentially to instigate emergency braking, etc. One example of radar technology that has been proposed for vehicular applications uses mono-static radar technology arranged to operate in the 77 GHz frequency range.
Integrated circuits (ICs) and/or communication units arranged to utilise radar technology require transceiver chips, or separate transmitter and receiver circuits to be located in close proximity to one another to improve the accuracy in determining a distance and bearing to the object. As a consequence, and particularly at typical radar frequencies where the transmit (and therefore receive) frequency is very high, such as in the 77 GHz frequency region, it is known that mono-static radar technology suffers from interference caused by poor isolation between the transmitted and received signals at these very high frequencies within the ICs or communication units.
FIG. 1 schematically illustrates known causes of interference effects in a high frequency communication unit 100. The high frequency communication unit 100 is illustrated with respect to a receiving operation, whereby an antenna 102 receives high frequency signals and passes them to a low noise amplifier 104. Depending on the receiver design and system requirements, the low noise amplifier is optional and can be also bypassed. The received and perhaps amplified high frequency signal 106 is input to a down-converting mixer 108, which down-converts the amplified signal 106 by multiplying it with a local oscillator (LO) signal 112 that is fed from an LO source 110. The output from the down-converting mixer 108 is a desired intermediate frequency (IF) signal 114, which is typically at a very much lower frequency, such that low-pass or band-pass filtering can be used to remove or attenuate undesired signals in the frequency domain. The IF signal may be a low frequency (LF) signal, a very low IF (VLIF) signal or even a DC (zero IF) signal. As shown, and particularly with high radio frequency (RF) signals, the signals may be undesirably radiated to nearby circuits/elements/transmission lines, etc. Thus, it is known that LO signals may radiate directly onto the receiver path, thereby causing interference. This interference is known as cross-talk interference or isolation cross-talk.
As a consequence, in order to reduce the level of interference that is radiated between internal circuits/elements/transmission lines, etc., many radar systems use ultra short transmit pulses to guarantee that the transmitter is shut-down (and therefore the transmitter oscillator signal is highly isolated from antenna) when the echo of the ultra short pulse is expected at the radar receiver. Alternatively, or additionally, radar systems may use spatially-separated antennas for the respective transmit or receive operation, with the spatially-separated antennas arranged to provide high isolation there between. It is also known that radar systems may use high-end circulators to reduce the interference effects.
Isolation to minimise cross-talk may therefore be achieved at high frequencies using high-end circulators or rat-race couplers 150, located between transmit and receive paths. A rat-race coupler 150 would typically provide less than 20 dB isolation between the two paths. Thus, for example, a radar transmit signal of 16 dBm at 77 GHz input to a rat-race coupler of 20 dB isolation would still leak −4 dBm of transmit signal 152 into the receiver chain. This level of leakage power will be significantly more than the desired receive signal. Hence, a significant portion of the transmitted signal couples into the receiving channel/circuitry. This undesired transmitted signal acts as an additional, unwanted signal in the receiver down-mixer circuitry, thereby creating further undesired down-converted signals in the intermediate or low-frequency/baseband circuitry that degrades the complete system performance.
High frequency mixer circuits are often based on the known Gilbert cell type. The Gilbert cell type is an active mixer that provides a conversion gain instead of conversion loss. However, the linearity of such active mixers is known to be limited. Thus, in a mono-static radar system, where the signal leakage may easily exceed −4 dBm, the mixer should still be able to operate in a linear mode with such a high leakage level. To achieve this level of linearity, the input referred 1-dB compression point, which is a measure for the linearity of the receiver, must be designed with sufficient margin compared to the maximum input power. As a rule of thumb, the compression point is calculated as: 10 dB plus the maximum power level. Thus, in the above example when the leakage level is −4 dBm, the desired 1-dB compression point is in the region of +6 dBm. Hence, the design of such extremely linear mixer cores requires high supply voltages and extremely high current densities in the transistors. As active Gilbert-cell mixers cannot support this combination of competing system parameters, a trade-off is often made, for example the output power of the transmitter is often reduced to lower the cross-coupled leakage into the receiver path. However, a lower transmitted power will reduce the signal-to-noise ratio (SNR) of the system, which in turn degrades the system performance. To overcome this problem, the cross-coupled signal into the receiver needs to be cancelled using an alternative approach.
In addition, the phase effects of radio frequency (RF) circuits, such as the Gilbert cell mixers, are difficult to be compensated for as implementing controllable phase shifter technology at such high frequencies is generally practically unrealizable due to cost, size and/or isolation performance constraints.